From the outset of the AIDS epidemic here in the U.S. public policy makers have had to consider its impact on the health care system. The little research conducted in the area suggests that health care costs incurred by persons with AIDs can be very high. Estimates of the spread of the disease suggest that hundreds of thousands have contracted it and most, if not all, will die, incurring very high medical and care costs in the process. The proposed study seeks to greatly expand our knowledge of these issues by building upon an evaluation of the Robert Wood Johnson Foundation's AIDS Health Services Demonstration to assemble a panel of patients with AIDs being served by demonstration projects in nine (9) communities across the country. Under the proposed project, this panel of approximately 1224 patients will be reinterviewed from 9 to 12 months after their initial evaluation interview and their hospital, clinic, physician and home care service use records will be abstracted covering at one to two years, generally from diagnosis to the followup interview or death. We will build upon the existing data collection instruments, the network of data collectors hired and trained by Survey Research Associates, Inc., and the core group of investigators at Brown, Harvard and collaborating institutions across the country. The overall goal of the project is to describe the pattern of medical and social service use of PWAs and to determine whether their needs for services are being met and whether their preferences for aggressive or palliative medical care change as their disease progresses. The specific aims include: 10 examining changes in patients' patterns of health and social service use; 2) determining the clinical, social and organizational factors that predict service use; 3) examining correlates of changes in patients' level of unmet need for services; 4) examining whether and how AIDS patients; preferences for aggressive versus palliative medical care change as disease progresses; and 5) examining the clinical and socio-demographic factors that predict patient survival. Analyses will focus on comparing the experiences of selected risk groups (women, IVDAs and gay men), patients served in hospital based versus community based settings, those early in the course of their disease versus those later in the course of the disease and those with varying clinical profiles (e. g. PCP versus KS). Multivariate analyses of the correlates health care costs will be performed for specific time periods post-AIDS diagnosis with the patient as the unit of analysis and using event history analysis. If funded, we intend to collaborate with a study being proposed by Arnold Epstein, M.D. by sharing interview and record abstract forms and by attempting to conduct parallel analyses whenever feasible.